Min-max model predictive control of nonlinear systems : a unifying overview on stability

D.M. Raimondo, D. Limon, M. Lazar, L. Magni, E.F. Camacho

Research output: Contribution to journalArticlepeer-review

123 Citations (Scopus)
3 Downloads (Pure)


Min-max model predictive control (MPC) is one of the few techniques suitable for robust stabilization of uncertain nonlinear systems subject to constraints. Stability issues as well as robustness have been recently studied and some novel contributions on this topic have appeared in the literature. In this survey, we distill from an extensive literature a general framework for synthesizing min-max MPC schemes with an a priori robust stability guarantee. First, we introduce a general prediction model that covers a wide class of uncertainties, which includes bounded disturbances as well as state and input dependent disturbances (uncertainties). Second, we extend the notion of regional input-to-state stability (ISS) in order to fit the considered class of uncertainties. Then, we establish that the standard min-max approach can only guarantee practical stability. We concentrate our attention on two different solutions for solving this problem. The first one is based on a particular design of the stage cost of the performance index, which leads to a Hoo strategy, while the second one is based on a dual-mode strategy. Under fairly mild assumptions both controllers guarantee ISS of the resulting closed-loop system. Moreover, it is shown that the nonlinear auxiliary control law introduced in [29] to solve the H8 problem can be used,for nonlinear systems affine incontrol, in all the proposed min-max schemes and also in presence of state-independent disturbances. A simulation example illustrates the techniques surveyed in this article. .
Original languageEnglish
Pages (from-to)5-21
Number of pages18
JournalEuropean Journal of Control
Issue number1
Publication statusPublished - 2009


Dive into the research topics of 'Min-max model predictive control of nonlinear systems : a unifying overview on stability'. Together they form a unique fingerprint.

Cite this